Full text of "Ignition system requirements and their application to the design of capacitor discharge ignition systems."

See other formats

IGNITION SYSTEM REQUIREMENTS AND THEIR

APPLICATION TO THE DESIGN OF CAPACITOR

DISCHARGE IGNITION SYSTEMS

Terrence Lyle WMliamson

NAVAL POSTGRADUATE SCHOOL

Monterey, California

THESIS

IGNITION SYSTEM REQUIREMENTS AND THEIR

APPLICATION TO THE DESIGN OF CAPACITOR

DISCHARGE IGNITION SYSTEMS

Terrence Lyle Williamson

Thesis Advisor;

R. W, Adler

December 1971

Appnx)VQ,d (^on. pubtic A.cXea6e; dUtAibatLon unlimitzd.

Ignition System Requirements and Their Application

to the
Design of Capacitor Discharge Ignition Systems

Terrence Lyle Williamson
Lieutenant, United States Naval Reserve
B. S. , Weber State College, 1965

Submitted in partial fulfillment of the
requirements for the degree of

MASTER OF SCIENCE IN ELECTRICAL ENGINEERING

from the

NAVAL POSTGRADUATE SCHOOL
December 1971

Ti^> < '-• / -^

ABSTRACT

Kettering ignition systems used on the majority of automotive engines can
no longer assure reliable ignition for high-output engines. The capacitor dis-
charge ignition, CDI, is a promising system to supersede the obsolete battery-
coil.

This study examines wave-front requirements at the spark plug for pro-
ducing ignition in the internal combustion engine and system characteristics
necessary for producing the wave-front. Arc requirements are described and
used to define CDI parameters.

A modified Kettering system which violates some basic ignition concepts
was replaced by a CDI system designed in this study. Its characteristics were
derived from arc requirements, not by aggrandizing the replaced battery-coil
parameters.

During performance tests, the CDI system exhibited superior performance.
It fired simulated fouled plugs and continued to produce an arc when pressurized
tx) 3 times the value at which the Kettering ceased to function. This improved
performance was accomplished with approximately the same stored energy
and less input power.

TABLE OF CONTENTS

L INTRODUCTION 8

n. IGNITION SYSTEM PARAMETERS 9

A. SPARK PLUGS 9

1. Fouling 9

2. Fouled Plug Simulation 11

B. SPARK PARAMETERS . 11

1. Gap Width 11

2. High Tension Voltage Requirements 13

a. Compression 13

b. Gap Spacing 14

c. Electrode Temperature 14

d. Speed and Load 14

e. Acceleration 16

f. Ignition Timing 16

g. Fuel-Air Ratio 16

h. Voltage Polarity 16

i. Electrode Condition 18

j. Overview of Voltage Requirements 18

3. Spark Duration 19

4. Voltage Rise Time 20

5. Energy Requirements 23

m. IGNITION SYSTEMS 25

A. KETTERING IGNITION SYSTEM 25

B. HITACHI IGNITION SYSTEM 29

C. PIEZOELECTRIC IGNITION 31

D. TRANSISTORIZED IGNITION SYSTEM — — — 32

E. CAPACITOR DISCHARGE IGNITION SYSTEM 33

1. Operation 35

2. CDI, Improved Ignition Characteristics 36

IV. IGNITION SYSTEM DESIGN REQUIREMENTS 39

A. BRUTE FORCE CRITERIA 39

B. GENERAL DESIGN REQUIREMENTS 40

C. SPECIFIC REQUIREMENTS FOR A SYSTEM WHICH IS TO
REPLACE THE HITACHI IGNITION SYSTEM 42

1. Maximum Voltage Requirements 42

2. Spark Plug Gap 42

3. Spark Duration 42

4. Storage Capacitor And Energy Requirements 42

5. Rise Time 43

6. Power Input Requirements 43

7. Sunmiary of Design Criteria 44

V. FEASIBILITY STUDY 45

A. HITACHI IGNITION EVALUATION 45

B. CDI SIMULATION 45

C. CONCLUSION — 48

VL GDI SYSTEM DESIGN 51

A. DG TO DG GONVERTER 51

B. DISGHARGE GIRGUIT AND MV GATE 55

G. SGR TRIGGER GIRGUIT 57

Vn. GDI SYSTEM BENGH TEST AND EVALUATION 62

VniXONGLUSIONS AND REGOMMENDATIONS 77

LIST OF ILLUSTRATIONS
Figure

1. SPARK WAVEFORM • 12

2. EFFECT OF COMPRESSION PRESSURE ON VOLTAGE
REQUIREMENTS 15

3. EFFECT OF ELECTRODE TEMPERATURE ON VOLTAGE
REQUIREMENTS 15

4. EFFECT OF ENGINE SPEED ON VXTAGE REQUIREMENTS 15

5. EFFECT OF ACCELERATION ON VOLTAGE REQUIREMENTS 17

6. EFFECT OF SPARK TIMING ON VOLTAGE REQUIREMENTS 17

7. EFFECT OF FUEL-AIR RATIO ON VOLTAGE REQUIREMENTS— 17

8. EFFECT OF IGNITION TIMING AND SPARK DURATION ON

ENGINE OUTPUT 21

9. ENERGY VERSES CAPACITANCE 21

10. KETTERING IGNITION SYSTEM 26

11. KETTERING IGNITION 26

12. HITACHI IGNITION SYSTEM 30

13. TRANSISTOR IGNITION SYSTEM 30

14. CDI BLOCK DIAGRAM 34

15. HITACHI MODEL 46

16. LISA SOLUTION FOR HITACHI MODEL 47

17. CDI MODEL 49

18. COMPUTER OUTPUT OF CDI SYSTEM 50

19. GDI BLOCK DIAGRAM 52

20. DC-TO-DC CONVERTER 54

21. DISCHARGE CIRCUIT AND MV GATE 56

22. UJT TRIGGER 59

23. FORWARD GATE CHARACTERISTICS 60

24. TURN-ON TIME CHARACTERISTICS 60

25. GDI SCHEMATIC DIAGRAM 61

26. HITACHI OUTPUT 65

27. HITACHI OUTPUT PHOTOGRAPHS 66-68

28. CDI OUTPUT 69

29. CDI OUTPUT PHOTOGRAPHS 70-71

30. SYSTEM WAVEFORMS 72-75

31. CDI SYSTEM POWER CONSUMPTION 76

I. INTRODUCTION

The ignition system currently used on the majority of automotive engines
has been improved only slightly since its introduction in 1914 by Charles F.
Kettering. This pioneer system, referred to as the battery-coil or Kettering
ignition is incapable of keeping pace with the demands put on it by todays engines.

This study deals with the requirements necessary to produce ignition of a
fuel-air mixture in an automotive engine combustion chamber. Ignition require-
ments are divided into arc characteristics for proper ignition and system require-
ments to produce this arc.

To evaluate the processes used to produce the arc, various ignition systems
are discussed. Their basic operation and characteristics are described to give
insight into ignition system requirements.

A specific, modified battery- coil ignition system is described and its char-
acteristics listed. This system is described in detail since a capacitor discharge
ignition, GDI, system is designed to replace it. Characteristics of both systems
were evaluated and the improved operating characteristics of the CDI system
noted.

Ignition requirements are established and are applicable to new ignition system
principles. These requirements are applied to the design of the system.

This study concludes by recommending the implementation of CDI as the
standard ignition for modern automotive engines.

IL IGNITION SYSTEM PARAMETERS

Abundant literature exists on ignition systems and on the spark character-
istics necessary to assure ignition of the explosive charge in the combustion
chamber. The difficulty arises in ferreting out what requirements are necessary
and applicable; the sources of information are not all in agreement on just what
parameters should be considered in the design of an ignition system.

This section defines ignition parameters and their relationship to proper
ignition.

A. SPARK PLUGS

The spark plug is that portion of the ignition system producing the arc that
ignites the fuel-air mixture. If the arc does not have proper characteristics,
ignition will not take place and misfiring will result. It is the duty of the ignition
system to supply the necessary voltage and energy to the spark plug.

Also, the spark plug is the common element in all ignition systems. Regard-
less of the driving source configuration, the plug produces ignition by an arc
occuring between electrodes. Gap configurations may vary, but arc generation
and its characteristics remain basically unchanged.

1. Fouling

Fouling can be attributed to metallic compounds found in combustion
deposits. These materials, accumulating on the insulator firing end, become
electrically conductive, under certain operating conditions, and can thus prevent
the ignition voltage from building up sufficiently to fire the plug.

Fouling is caused by many factors such as: engine make and model,
engine power utilization, spark plug design and heat range, anti-knock additives
and other fuel additives, and oil consumption. It occurs due to accumulation of
deposits under low temperature (low output) or high temperature (high output)
conditions.

Dry, fluffy black carbon deposits result from overrich carburetion,
excessive choking, or a sticking manifold heat valve. Low ignition output can
reduce voltage and cause misfiring. Excessive idling and slow speeds under
light load also can keep spark plug temperature so low that normal combustion
deposits are not burned off.

Deposits accumulating on the insulator are by products of combustion
and come from the fuel and lubricating oil, both of which today generally contain
additives. Most powdery deposits have no adverse effect on spark plug opera-
tion; however, they may cause intermittent missing under severe operation
conditions, especially at high speed and heavy load. Under these conditions
the powdery deposits melt and form a shiny yellow glaze coating on the insulator
which, when hot, acts as an electrical conductor. This allows the current to
follow the deposits instead of jumping the gap.

The average driver cannot operate in a range that will best prevent
fouling. He is likely to be subjected to both types of fouling since he will drive
under low output while in city traffic yet in a high output situation on the express-
ways. This requires the ignition system to fire a fouled plug under all engine
operating conditions.

2. Fouled Plug Simulation

As discussed above, a fouled plug will present a conductive shunt

pjath for the ignition current. A high voltage noninductive resistor connected

2 3
from the spark plug to ground may be used to simulate a fxiled plug. * This

resistance is usually in the range of 0. 5 to 1. Mi\. The 1 Ma. test is

intended to simulate system performance with a fouled plug and is an industry

standard test.

B. SPARK PARAMETERS

To ignite the fuel-air charge in an internal combustion engine, the spark
must meet certain criteria. The basic parameters effecting the spark are gap
width, high tension voltage, spark duration, rise time, and energy.

Figure 1 is the waveform observed on an oscilloscope placed across the
gap of a spark plug. Areas 1 and 3 iqjresert the ignition voltage rise time and
arc sustaining voltage duration respectively. The arc is struck at point 2 and
extinguished at point 4.

1. Gap Width ^'^'^*^

Cycle-to-cycle variations in ignition consistency is related to the
condition of the gap. The plug location and purging of the gap are important.

The gap must have a minimum spacing to enable the arc to transfer
adequate heat energy to the fuel-air mixture. The mixture has a natural tendency
to quench or cool everything within its path. Wetted spark plug electrodes pro-
duce boundary layers of fuel-air ratios too rich to ignite.

A wide gap will enhance the circulation of mixtures of ignitable
ratios within the gap area. A lean mixture, with greater molecular spacing of

>
o
5

o
a.

fuel and air particles, requires a wider gap in order to allow adequate quantities
of the mixture within the proximity of the spark for the necessary heat transfer to
Initiate combustion.

Widening the gap increases the capacitive energy delivered to the gap.
This is a result of the increased voltage required to initiate the arc between the
electrodes. Since E = |CV , the additional energy available will be proportional
to the square of the additional voltage required.

The gap is typically set at the minimum value that provides smooth
engine idle. Basically, gap width should be as large as possible but not so large
that the ignition harness and distribution system will not handle the voltage required
to create ionization. Allowances must also be included for gap growth.

In the past, gap widths have been limited by available ignition voltages.
This problem has been diminished and large gap widths, up to 0. 050 in. , may
now be recommended. The limiting factor now is the voltage breakdown of the
ignition system components. The gap width should be limited so that the maximum
voltage required under the worst conditions is approximately 22 kV.
2, High Tension Voltage Requirements

The voltage required to cause arc-over is dependent on engine design

and operating conditions as well as spark plug geometry. Variations from 4 to

3 6 7 8 9
20 kV. for various engine operating conditions are not uncommon. » » » ♦

a. Compression

The variable which is usually considered first, due to the tradi-
tional method of bench testing, is compression pressure. While the absolute
value of sparking voltages will vary somewhat depending on the type of fuel.

moisture content and voltage source, this is basically a linear relationship,
as indicated in Fig, 2, with the voltage required increasing as pressure sur-
rounding the gap increases. However, because of the nonuniform electric
field gradient within the plug gap, the breakdown voltage does not follow
Paschen's Law exactly.

In general, Paschen's Law states that the voltage required
to jump a given gap in a uniform field is dependent only upon the product of the
gas pressure and the electrode spacing. Increased gap size or pressure results
in less breakdown voltage than predicted by Paschen's Law.

b. Gap Spacing

Other factors being equal, sparking voltages will increase
directly with gap spacing within the normal range of usable settings as shown
in Fig, 3, The discussion on Paschen's Law above also applies here.

c. Electrode Temperature

Temperature has a marked effect on voltage requirements.
The lower the temperature, the higher the voltage required to cause arc over.
This effect is also shown in Fig, 3.

d. Speed and Load

The effects of speed and load in a typical 4-cyle automotive
engine are illustrated in Fig. 4, The slight decrease noted at high speeds can
be attributed to increased spark plug electrode temperatures and decreased
compression pressures which occur as the engine's breathing efficiency decreases.

Voltage
Required

Fig. 2.

Compression Pressure

Voltage
Required

Electrode Temperature

Wide Open
J Throttle

Roa d Load

Speed

Fig. 3

Fig. 4

e. Acceleration

Sudden, wide-open-throttle acceleration causes rapid but
temporary rises in voltage requirements as shown in Fig. 5. This increase is
attributed to the rapid increase in pressure. The effect here is greater than
that due to temperature, since the spark plug electrodes have not had time to
heat up.

These sudden voltage. increases are transient in nature and
explain why misfiring is often encountered first during periods of rapid accel-
eration,

f. Ignition Timing

The typical effect of spark advance on voltage is illustrated
in Fig. 6. Advancing ignition timing lowers voltage requirements because the
spark plug fires at a lower pressure, and the electrodes are hotter because of
less charge cooling and higher flame temperatures.

If the spark is retarded past top dead center, requirements
decrease as compression at the point of ignition drops, and power and temper-
ature are reduced.

g. Fuel-Air Ratio

Lowest voltage requirements will be observed at the stoichi-
ometric ratio as shown in Fig. 7. Leaning of the fuel charge has the greatest
effect on voltage requirements, although the overall effects can be considered
negligible in the normal range of fuel-air ratios.
h. Voltage Polarity

Voltage polarity, commonly called coil polarity, is an often
overlooked yet important factor. It must be considered because on all conventional

Voltage
Required

Acceleration

Fig. 5

Speed

T. D.C.

Voltage
Required

Fig. 6

Adv. Ret.

Spark Timing

Voltage
Required

066

Fig. 7

Lean Rich

F/A

spark plug designs, the center electrode operates considerably hotter than the
ground electrode. Electron theory states that electrons move more readily from
the hot to the cold electrode than the inverse. Therefore, the voltage applied to a
plug must rise in the negative direction in order to produce ionization at minimimi
voltage.

If the plug has reverse polarity with respect to that defined
above, the firing voltage is greater. In some instances, the difference may be
a few thousand volts.

i. Electrode Condition

Sharp or pointed electrodes concentrate the gap ionization
by increasing the electric field gradient. Therefore, spark plugs can be expected
to require progressively greater voltage as the sharp corners of the electrodes
erode away and become rounded in normal service.

Fouling deposits do not influence the arc-over voltage, unless
the deposits are within the gap area, which is seldom the case.

j. Overview of Voltage Requirements

The maximum available voltage from the coil should not
exceed 30 kV. as any voltage higher than this can produce undue strain on the
ignition harness, distributor, and spark plugs. At any voltage exceeding 30 kV.
the spark plug insulation can flashover either internally or externally, and
similiar flashovers can occur within the distributor either from the cover
electrodes to ground or between electrodes. These flashovers can form carbon
paths that once started can seriously down-grade engine performance.

Manufactures of electronic ignition systems have advertised
voltages as high as 60 kV. , pursuing the theory that if some is good, more is
better and the more voltage, the better the system. Actually, under worst
conditions, an engine should not require more than 22 kV. across the plug to
establish ionization. Under normal conditions, the peak voltage is consider-
ably lower than this. The objective of an ideal system would be to produce the
required voltage at any engine speed while maintaining the proper energy for
the required length of time.
3. Spark Duration

Optimum spark duration for ignition of the fuel-air mixture has not
been determined. Values range from a microsecond to thousands of micro-
seconds.

At present, spark duration requirements are evaluated on test
engines where the duration that provides the highest engine output is considered
optimum. This procedure does not demarcate the requirements for ignition of
the combustible mixture, but rather the deficiencies that exist in combustion
chamber design.

For ignition to take place there must be a combustible mixture
between the spark gap. Longer spark durations have a higher probability of
igniting a mixture that is not homogenous throughout the combustion chamber.
Long arc durations give sufficient time to permit the fuel charge to come within
the gap area.

Ignition timing has a large control over the turbulences that exist
in the combustion chamber. Figure 8 illustrates the effect of ignition timing

and spark duration on engine output. Analyzing Rg. 8 more closely, it is seen
that spark duration has limited effect on engine output if the specified engine
timing is used, therefore, overadvanced timing is avoided.

Practically all of todays automotive engines use the battery-coil
ignition system with spark durations typically 1,000 to 2,000 usee. Current
references on ignition systems recommend long duration times of 500 usee, or
.joaore, where possible. However, an example of the ability of the capacitive
portion of the discharge to initiate combustion has previously been demonstrated
with the piezoelectric ignition system developed by Clevite Corporation in the
early 1960's. The entire pulse width of the system was only 860 nsec. which
is less than the rise time of a conventional magneto.

Theoretical discussions on ignition systems indicated that a system
with such short duration would not fire the mixture, but the piezoelectric
system does fire the mixture and very well too.

Due to combustion chamber variations from engine to engine, and
even from cylinder to cylinder in the same engine, it is recommended that
the spark duration be at least 100 to 200 usee, in duration. If the combustion
chamber and engine design are satisfactory this duration will be sufficient to
ignite the fuel-air mixture.

4, Voltage Rise Time

From examining various ignition systems, their past history and
theoretical arc considerations, it appears that voltage and energy are not the
only criteria of ignition system operation. Many of the systems that are capable
of firing fouled spark plugs exhibit shorter rise times than the battery-coil

Specified

12 gsec

Ret. Adv. — ^

Ignition Timing

Fig. 8

20--

Energy
(mJ)

60
Capacitance ipF)

100

Fig. 9

system. It is concluded that rise time must be included in ignition system
evaluation and design. For example, in referring to the piezoelectric system
mentioned earlier, the rise time for this system was not measured exactly but
was less than 10 nsec. Some investigations indicate it may be as short as one
nsec. This short rise time system ran an engine six times longer between plug
replacement than a longer rise time magneto; yet, in some circles, the magneto
is thought to be the ultimate.

Rise time is defined as the duration required for the voltage to
build up and fire a spark plug. If this time is short, there is less opportunity
for the energy to be dissipated in carbon deposits, moisture, and other partial
conducting paths. Also, a short rise time is more effective in ion formation.

Paschen's Law must again be evaluated in the realm in which it was
written, namely that the electric field is uniform and that the voltage is slowly
applied. For a short rise time pulse, the gap voltage may be reduced below
the value considered normal by at least 15 percent.

Shorter rise times have the disadvantages of larger radiation losses
(increased radio interference), increased requirements on the system to pre-
vent crossfiring, and more chance of developing unwanted carbon paths.

An optimum ignition pulse would have sufficient rate of voltage rise
to permit firing of heavily fouled plugs without the need for large total pulse
energy. The rate of rise should be consistent with the voltage breakdown of
the ignition system components. A rise time between 10 and 30 usee, should
prove adequate for most cases.

5. Energy Requirements

Total energy is important, in that a certain minimimi energy is
required for ignition, but the required energy depends to an extent on the rise
time and pulse width of the arc. Energy levels higher than necessary to account
for the variables are detrimental to spark plug life. Since the energy level
required for the standard mixture may be as low as 0. 002 mJ. , typical igni-
tion system energies are higher than necessary. In general it is considered
that 1 mJ. is sufficient to produce ignition of the fuel-air mixture.

The energies mentioned above are those required to raise a small
amount of mixture to combustion temperature. Further, this quantity of energy
is actually a very small part of the total energy a system must have. Literature
on system requirements list system energies from 10 to 40 mJ. These large
energy requirements, compared to that required for combustion, are due to

system losses and capacitances. If losses are neglected, energy requirements

2
reduce to E = ^CV . Assuming that under all operating conditions 22 kV. is

sufficient to arc across the gap. Fig. 9 shows the energy required to overcome

system capacitance. If a short rise time system is used little energy will be

dissipated prior to the arc. Once the arc is struck, the capacitive energy is

released rapidly in the leading section of the arc; inductive energy is released

slowly increasing arc duration.

Energy requirements should be held to a minimum. If excess energy

is delivered to the arc, no improvement in combustion is noted, furthermore, gap

erosion increases. Designers of ignition systems must consider the fact that

only a small, insignificant part of the total energy is required to produce
combustion. The majority of system energy is used to assure that the required
voltage will be developed.

in. IGNITION SYSTEMS
The contents of this section describe the operation and characteristics
of some of the ignition systems in use today in the automobile. Two ignition
systems that are not in wide spread use are also discussed. This section
covers both the non- electronic and electronic systems.

A. KETTERING IGNITION SYSTEM

Since 1914, automakers have used the Kettering or inductive ignition
system — a battery, ignition coil, and cam-driven mechanical switch. Most
of the automobiles sold in the U. S. come equipped with this system. Fig. 10
is a representative schematic diagram of a typical Kettering system.

Referring to Fig. 10, when the ignition switch SWl and the cam operated
contacts SW2 close, current will flow through primary P of ignition coil T,
building up magnetic flux. The current will reach a maximum value limited by
the resistance of the primary. As the cam rotates, contacts SW2 are separated
by the cam lobes, interrupting primary current. The distributor contact
capacitor CI suppresses contact arcing and forms an oscillatory circuit with
the equivalent primary inductance of T.

Interruption of the primary current causes the flux in the ignition coil to
collapse. The collapsing flux self-induces a voltage in the primary and by
mutual coupling induces a voltage in the secondary of T. The secondary volt-
tage is related directly to the primary voltage as the ratio of the nimiber of
turns in the secondary to the number of turns in the primary.

Fig. 10

Contacts

Fig. II

The high voltage causes a spark to jump the small gap between the rotor
R and the distributor cap insert with which the rotor is aligned, thereby firing
the desired spark plug, SP, connected to the insert by a high tension cable.

Figure 11 is also a schematic of the Kettering system, but contains
some of the distributed components that are often overlooked. Cs is the
capacitance of the secondary of T and the high tension leads. Rp and Rs repre-
sent the internal primary and secondary resistances, respectively, of the
induction coil and Cc is the contact capacitor.

If a perfect ignition coil is assumed which is free from loss due to resist-
ance, radiation, and dielectric hysteresis, the coil output energy will be equal
to the coil input energy. The input energy in an ignition coil is related to the
primary inductance and primary current by the equation:
W =■! Lplp^

where: W = energy stored in primary
Lp = primary inductance
Ip = primary current.

The current in an inductive circuit is a function of the time the circuit is
energized. If the ignition, coil energy is not to decrease by more than 10 percent
at some high speed value, the time constant, T = L/R, of the coil primary
circuit must be one-third of the high speed primary circuit actuation time
(contacts closed for three time constants).

It is necessary to evaluate the total energy required in the primary of a
coil. Assuming an ideal coil, to elevate the secondary to spark plug firing

voltage,

Wt = ^ [Cs + (Np/Ns)^Cs] Vs^ (1)

where: Wt = total energy in system

Cs = total secondary capacitance

Vs = secondary voltage

Ns = turns on coil secondary

Cc = contact condenser capacitance.

From the above analysis a typical standard ignition system requires a

primary current of- 16 amperes. This current must flow through the contacts,

however, it is impractical for a contact set to handle this much current on a

continuous basis. Therefore, current limitation forces conventional systems

to operate at a maximum of about 5 amperes with a resulting decrease in high

speed performance.

12
The disadvantages of the Kettering system are:

1. The large value of current being interrupted by the contact-
breaker points, cause excessive erosion.

2. The moving arm of the contact-breaker tends to bounce at high
speeds, thus shortening the time the points are closed. Point bounce reduces
coil output and also increases point wear.

3« A substantial reduction occurs in the output voltage with increas-
ing engine speed,

4. The system is highly inefficient at low engine speeds due to the
high current.

5. The system has a long voltage rise time resulting in poor
performance when spark plugs become fouled.

B. HITACHI IGNITION SYSTEM"^

Examining Fig. 12, it is seen that this system is a modification of the
Kettering ignition system described previously. The spark voltage is developed
in an identical manner and energy requirements remain the same. The system
also has the same disadvantages.

By using the Hitachi ignition system, a 4-cylinder, 4-stroke engine can
have the proper ignition sequence without the need for a distributor, and requires
only the ignition coil and breaker-plate. The high tension cables are connected
directly to the spark plug from the induction coil. By using the dual system as
shown, dwell time is doubled over that for a Kettering system on a 4-cylinder
engine. This improves high speed performance since primary current will
have a longer time to build-up to the design value.

The system has some major drawbacks that require discussion, dealing
with the way in which the voltage is delivered to the plugs. Notice that two plugs
are fired simultaneously in series with respect to the induction coil termination.
This is permissable, since one plug fires on the power stroke while its mate
fires on the exhaust stroke.

The disadvantage is that higher potential must be developed to produce
arcs in two plugs in series instead of just one. Since one plug is firing on the
exhaust stroke, the potential required will be much less than that for the plug
firing on the compression stroke. The main disadvantage, however, is that one
plug is being fired with reverse polarity. As mentioned earlier, a plug fired with
reverse polarity requires a few thousand volts more .

Contact- Breaker

I
i

Fig. 12. Hitachi Ignition System

► O-

Fig. 13, Transistor Ignition

This system must still adhere to requirements that ignition voltage be
between 25 kV. and 30 kV. To compensate for the need for higher ignition
voltage, the spark plug gap is reduced slightly to lower the arc-over potential.
This reduction in gap width will of course decrease the arc area available for
the fuel-air mixture to circulate.

The reason for considering this particular ignition system is that an
electronic ignition system was designed in this study to replace it.

C. PIEZOELECTRIC IGNITION

The piezoelectric ignition system has not been commercially used on
production engines. Its introduction here is to inculcate the point that rise
time, being of little concern to ignition system designers until recently, should
play a larger part in the design of systems and to point out that extremely long
arc durations in the thousands of microseconds are not required to produce
combustion of a homogenous fuel-air mixture.

This ignition system derives its name from the piezoelectric generation
of electricity in a crystal structure when pressure is applied. System operation
is exceptionally simple in theory. The potential difference generated by a
crystal, a stack of crystals in series, when struck sharply by mechanical
means is applied to the spark plug. The voltage rise is extremely rapid and
the energy delivered to the arc is strictly capacitive, therefore, the arc is of
short duration. As mentioned before, the characteristics for this system are
in the nanosecond range.

Older theoretical discussions indicate that an ignition system with this
short rise time and pulse width can not fire the mixture. This system has

run an engine six times longer between plug replacement than has a magneto with
its long rise time and arc duration. It can start an engine after the spark plug
has been soaked in water and put into the engine dripping wet.

D. TRANSISTORIZED IGNITION SYSTEM

The use of a transistor switch was one of the first attempts to use semi-
conductors to improve the Kettering ignition system. The transistorized
system is essentially identical to the conventional one except for the addition
of transistor Ql, Fig. 13. The discussion of the Kettering system, section
IIL A. , is applicable.

The difference is that the primary current is now switched on and off by
a transistor instead of the contact-breaker. The contact condenser is also
eliminated with only the small collector to emitter capacitance in the primary
circuit. The points control the base current thus turning Ql on and off. The
small current in the base does not cause the points to erode as rapidly as in the
Kettering system. A light or magnetic sensing device can be connected to the
base of Ql to turn it on and off, thus eliminating the points entirely.

By eliminating the points, a larger primary current can be used to
improve high engine speed performance, the higher currents being obtained
by reducing the primary inductance. This usually results in an increased
turns ratio. Standard Kettering ignition systems usually have a turns ration

of 100:1 while the transistorized systems have a much higher turns ratio,

14
often in the vicinity of 250:1 to 500:1.

Although the transistor ignition system produces a more constant voltage
throughout the engine speed range, it still has a long rise time and puts a large
demand on the battery and charging curcuit due to the requirement for increased
primary current. Under starting conditions, this system may not perform^
satisfactorily. During cold weather operation it is often inferior to the battery-
coil ignition system.

The transistorized system was used on some production automobiles. It
was soon discontinued since the small improvement in ignition high speed per-
formance did not warrant its additional cost. Cold weather starting reliability
was also less than the standard system,

E. CAPACITOR DISCHARGE IGNITION SYSTEM (CDI)

Capacitor discharge ignition systems have been on the market for a
number of years. The CDI system was used as an electronic ignition long
before transistorized ignitions were introduced. Figure 14 is a block diagram
of a typical CDI system. In the early systems, the dc-to-dc converter was of
the mechanical vibrator design and the gate was a thyratron tube. It was unre-
liable, yet produced superior ignition.

The CDI system remained obscure until the advent of semiconductor com-
ponents. With the introduction of power transistors and the SCR, the dc-to-dc
converter was easily produced using blocking transformers and the thyratron
was replaced by its counterpart, the SCR, This ignition is challenging the
Kettering system as the one for todays high performance engines, in fact, one

^ Vi ^

(

^"v^Y-r^

Timing
Source

D-D

rter

I c
O o

Q o

E
o

o
o

O
O

auto manufacture is using a GDI system as standard equipment on one of its
1972 models. It had been offered as an option at a substantial price prior to
this time.

1. Operation

There are two ways in which to use the ignition coil to produce the
high voltage pulse. First is the rate-of- change of current, or inductive mode.
This is the mode in which the Kettering system operates. Second is the trans-
former mode. In this mode the coil acts only to transform a low voltage to the
high voltage required. It is this second mode in which the capacitor discharge
system functions.

The dc-to-dc converter increased the low battery voltage to an inter-
mediate level of a few hundred volts. The output charges the capacitor, referred
to as a storage capacitor, to the intermediate voltage. At the proper time for
ignition, the trigger circuit opens the gate which in turn connects the storage
capacitor across the ignition coil. The capacitor voltage is then multiplied by
the transformer's turns ratio to produce the high voltage for ignition. By using
the transformer mode a much shorter rise time can be developed. The ignition
coil can be designed to have low inductance and thus act as a pulse transformer.
This can result in an ignition system with an extremely short rise time.

Ignition coil primary pulse duration is shorter than the response

time of the secondary. This means that even a capacitor storage type system

15
does store some energy in the transformer magnetic field momentarily. If it

was not for this magnetic field storage, the capacitor energy would be delivered

very rapidly to the arc resulting in a very short duration pulse. It is the energy
stored in the inductance that extends the arc duration since, as mentioned
earlier, inductive energy is released slowly.

Storage capacitor Cs must supply energy for the same reasons as dis-
cussed previously in regard to inductive systems.

Conduction of the gate connects Cs through the reflected impedance of
the transformer to the secondary capacitance of the ignition system. Equating
the energies on both sides of the perfect transformer and solving for Cs yields:

Cs = Vs ^ Cd

(2)

Vp^- Vs^ (Np/Ns)^

where: Cs = energy storage capacitance

Vs = secondary voltage

Vp = Cs voltage before SCR conducts

Cd = secondary distributed capacitance

Np = primary turns, ignition coil

Ns = secondary turns, ignition coil.

- The above relationship holds for all values of voltage and capacitance when

losses due to imperfect transformer action are neglected.

2. CDI, Improved Ignition Characteristics

The CDI system stores the energy required for ignition in a capacitor.

This form of storage has the advantage that once sufficient energy has been

accumulated in the capacitor, no more energy is consumed by the system until

the capacitor has to be recharged for the next firing. This means that the

system will draw only the energy it needs and therefore, current requirements

will vary as engine speed. In terms of energy requirement verses speed, the
CDI system has a higher efficiency and improved utilization of energy drawn
from the battery.

The CDI system, with its shorter rise time, provides better perform-
ance in firing fouled plugs. P. C. Kline of Delco-Remy reports that their

experience with the Delco CD system shows plug life 4 to 5 times longer than

9
with conventional ignition. FIAT, on their sports cars, used the CDI in

extending the thermal range of the spark plug so that a cold spark plug can be

used for highway operation, and at the same time avoid misfirings due to cold

9
fouling at low speeds.

One definite benefit is improved starting, particularly in damp

weather, or in very cold weather. CD equipped engines tolerate carburetor

flooding and other problems that cause starting difficulties.

Champion Spark Plug Company studied the effect of capacitor discharge

ignition on electrode erosion. Champion noted that spark plug gap growth was

much less when using the capacitor discharge ignition system with the fast rise

time and short arc duration. In fact, the spark plug gaps from the capacitor

discharge system actually decreased slightly due to a light deposit build-up.

Gap growth measurements are not a true indication of overall deterioration.

Center electrodes from the conventional system were round while the electrodes

from the capacitor discharge system still had relatively sharp edges. Sharp

edges are desirable since they reduce the voltage required for arc production.

The trigger circuit for a CDI system can be designed using a light or

magnetic sensor instead of the customary contact-breaker.

An ignition system having the above characteristics, if its electronics
were properly designed, would require a great deal less maintenance than the
standard battery-coil ignition. Since the contact-breaker assembly could be
eliminated, the only wear would be in the shaft bearings of the distributor
assembly. Once ignition timing was initially set, it would not need resetting
unless major maintenance was necessary on the distributor. The need to re-
move spark plugs for cleaning, regapping, or replacement is greatly reduced,
thus, greatly extending plug life.

The CDI system has the characteristics that are badly needed on todays
automotive engines. This system, if properly designed, could be the long-needed
replacement for the 1914 Kettering ignition system.

IV. IGNITION SYSTEM DESIGN REQUIREMENTS

The contents of this section define the requirements one must consider in
the design of ignition systems. Presented are current ideas on ignition system
design. Applying these concepts will result in systems which appear to be
better solutions. However, some procedures, if followed, mi^t give only a
brief reprieve from the problem the system was to eliminate before introducing
problems of its own.

A. BRUTE FORCE CRITERIA

If some is good, more must be better ! R. G. Van Houten and J. C.
Schweitzer of Delta Products states, "Any new ignition system must meet the
following requirements:"

1. "Output energy levels should exceed present levels by substantial
margins. A new system should be able to develop energies of 40 milliwatt-
seconds minimum, and be easily controlled to set this level higher if necessary.
The energy output and voltage levels should remain constant, over an rpm range
of 8, 000 to 10, 000 on eight- cylinder engines. "

2. "As rapid a voltage rise time as possible."

3. "It should be low cost and designed for high volume production. "

In the description of another system, the following statement is made, "It
has been pretty well established that a minimum of 30 milliwatt- seconds of

energy is required at the spark plug in modern ignition systems. CI has been

chosen to give 80 milliwatt- seconds, allowing ample reserve energy."

Again the system designer thinks in terms of brute force, "... we find
that it takes about 40 kV. to operate the spark plugs. This 40 kV. should be

considered a minimum requirement. To assure complete combustion, this

14
value should be exceeded if possible. " (Manufacturers of electronic ignition

systems have advertised voltages as high as 60 kV. )

Referring to previous discussions on ignition system requirements, energy
requirements are related more to system losses than to the energy required to
ignite the fuel-air mixture. Energy requirements necessitate careful system
evaluation and not the setting of a blanket value. The energy is not held in
reserve as mentioned above, but any excess energy, over that required to
ignite the mixture and compensate for system losses, is dissipated in the arc
and leads to excessive electrode erosion. Some German aircraft during World
War II used a high energy CD system to facilitate cold starts. Because of the
high energies involved, spark plug life was only 25 hours.

Under normal operation, a spark plug requires only about 4 to 8 kV. to
produce an arc. However, since the engine will be operating under various
load requirements, a voltage of 22 kV. is considered ample.

B. GENERAL DESIGN REQUIREMENTS

Following is a number of design criteria to be considered in the design
of ignition systems.

1. Use as short a voltage rise time as practical, not necessarily as
short as possible. With a sufficiently short rise time, an ignition system can
more readily fire fouled plugs. In selecting the upper limit on rise time,
capacitance loading, corona loss, and insulation failure become of paramount
importance.

2. A new ignition system must be more reliable than the system it
replaces. Reliability includes the time and cost of maintenance.

3. If the system is not original equipment, installation should require
a minimal change in components or wiring.

4. Input power should vary as engine speed.

5. Use energy levels only sufficient for operation.

6. Gains should be made by well known ignition practices related to
voltage, namely:

a. Keep the capacitance of the ignition leads as low as possible
by keeping them away from metal parts.

b. Reduce secondary series resistance to that required for radio
suppression.

c. Use short leads.

d. Reduce corona losses, and hysteresis of coils and capacitors.

e. Develop only sufficient voltage to assure that an arc can be
produced at all engine load and operating conditions. A small voltage reserve
may be applied, but should not be overdone.

7. If the system is designed to replace an existing one, leave the
original system intact so that it may be readily reconnected in case the new

system fails.

8. Be able to operate at temperatures in the engine compartment,
preferably as high as 250° F. , to permit installation directly on the fire wall,

C. SPECIFIC REQUIREMENTS FOR A SYSTEM WHICH IS TO REPLACE THE
HITACHI IGNITION SYSTEM

Figure 27 is a table of the characteristics of the Hitachi system. The
design parameters for the CDI replacement system conform to, or are im-
provements on, the Hitachi parameters.

1. Maximum Voltage Requirements

The ignition system was designed to supply 25 kV. optimum but less
than 30 kV. to protect high tension components.

2. Spark Plug Gap

The CDI, due to shorter arc duration, uses a wider gap than the con-
ventional system. The Hitachi system gap is set at 0. 6 to 0. 7 mm. — for the
CDI system the gap was widened to 1. mm.

3. Spark Duration

Arc duration was selected as 200 usee, to assure consistent ignition.
The storage capacitor was varied until a value was established that optimized
between spark duration and energy required.

4. Storage Capacitor and Energy Requirements

By equating energy on both sides of the ignition coil, equation (2)
was derived. This was used in calculating the energy storage capacitor,
which in turn was used to establish how much energy was stored. The param-
eters used were:

Cs = 0. 2 uF. from equation (2)
Cd = 20 pF,
Vp = 600 V.
Np = 380 turns
Ns = 15,000 turns.

The stored energy calculated, 36 mJ. , is close to the 30 mJ. standard

discussed earlier.

5, Rise Time

Rise time of the voltage pulse applied to the high tension circuit will
not be shorter than 10 usee, so that corona and radiation loss will be held to a
minimum. Rise time is to be no longer than 30 usee, to reduce energy loss and
high speed timing error.

There is no direct way to control rise time in the design of this
system since the original ignition coil is used. By the use of a capacitive dis-
charge through the coil, response of the coil was improved resulting in a
shorter rise time.

The use of the original ignition coil, rather than a specially designed
transformer, was one of the factors evaluated.

6. Power Input Requirements

In section IV. B. 4. , a storage energy of 36 mJ. per ignition pulse was
calculated. Maximum input power is required at maximum engine speed. The
ignition system must fire a 4-cylinder, 4- stroke engine, delivering peak bhp.
at 8500 rpm. A design margin of 1500 rpm. is included yielding a maximum
design rpm. of 10, 000.

At maximum rpm. the ignition system requires, for storage capac-
itor energy, 12 watts assuming 100 percent efficiency. Taking into consideration

the efficiency of the dc-to-dc converter, the power to operate the trigger
circuit, and sufficient current to keep the contact-breaker clean, an efficiency
of 70 percent is assumed. Thus, system power input is less than 17 watts at
an engine speed of 10, 000 rpm,

7. Summary of Design Criteria

Listed below is a summary of the requirements considered as goals in
the design of the Hitachi replacement:

1. Maximum engine speed is 10,000 rpm.

2. Maximum high tension voltage between 25 to 30 kV.

3. Arc ionization duration is 200 usee.

4. Rise time, 10 to 30 usee.

5. Capacitor storage energy, 30 to 40 mJ.

6. 15 to 20 watts of power consumption from a 12 V. dc. system,

7. Use the original ignition coil,

8. Operate over a temperature range of to 80*^ C.

9. Design for limited area installation.

V. FEASIBILITY STUDY

Before an attempt was made to design a CDI system, it was realized that
a pre- evaluation should be considered to test the feasibility of designing the
circuit around the parameters listed earlier. To accomplish this task, the
IBM 360/67 digital computer was used in association with the IBM circuit
analysis program LISA. During the computer evaluation, circuit parameters
were varied to determine their effect on system performance.

To have a means of comparison, the Hitachi ignition system was evaluated,
followed by the CDI evaluation.

A. HITACHI IGNITION EVALUATION

The Hitachi system was modeled as shown in Fig. 15. Initial condition
primary current was calculated from the time the ignition points are closed.
The computer solution commenced at the time of point opening.

LISA uses linear nodal analysis techniques. Since the characteristics of
an arc are highly nonlinear, a complete solution was unobtainable. The spark
plug was replaced with a load resistor, then the computer gave an indication
of the system rise time and available ignition voltage. Fig. 16 is a table con-
taining data from the solution.

B. CDI SIMULATION

Of interest was the question, could a capacitor discharging through the
primary of the Hitachi ignition coil produce the desired results listed earlier?

•-vVNA-J

u
a:

c
o

•o

a
■o

—

•*-

Q.
II

Q. M

♦-

k.
k.

a>
o

c
c

c
o

a
O

1-
a>

o
O

•*-

(0
<D

•o

»-

I-

>»

„_

i-
O

T3
O
O

C
O

o
®

■*-
a
o

o
o

£3.

o
o

X
ill

■-^

rf'

<£)

t-l

CQ
CO

•iH

o
o

w
o
<

to
•1-1

o
o
o

Q
U

To answer the question, the GDI system ^\as modeled as shown in Fig, 17. Rg
and impulse driver Vg were used only to implement LISA. Again, due to the
nonlinear properties of the arc, the spark plug was replaced with a resistance.

In the analysis it was assumed that the initial condition voltage of the
storage capacitor was not affected by engine rpm; this is valid as long as the
charging source has a sufficiently low impedance. Therefore, solutions were
not obtained for varying engine speeds. However, the storage capacitor value
was changed along with other circuit components to evaluate their effects on
ignition output.

The computer output verified the selection of a 0. 2 uF. storage capacitor.
A portion of the solution is shown in Fig. 18. Referring to Fig. 18, the rise
time falls between the limit set but the output available voltage is slightly
higher than the 30 kV. upper limit.
C. CONCLUSION

The computer analysis demonstrated the superior operation of the CDI
systen o\er the battery -coil system even when the same coil was used for both
applications. The solution indicates that a system could be designed to meet
stated specifications.

On the basis of the computer output, it was decided to proceed and to
design a CDI system to replace the Hitachi.

V\AA

•a
o

WNA

u
QC

■♦-

•—

<*-

•*-

■o

o
E

c
o
o

"■•

tl
a.

II
tn

-J

2
II

Q
O

•

■^

•»-

(/)

■o

c
o

-J

■o

(l>

o
o

Q-

•a

Output
Voltage

psec

Primary
Current

600

Voltage

300

|0 20 30

Fig. 18. GDI, Computer Output.

p sec

psec

VI. CAPACITOR DISCHARGE IGNITION SYSTEM DESIGN

System design was based on the requirements presented in section IV. B. 7.
which were adhered to except when situaticns developed that required modifica-
tions.

Reviewing Fig. 12, one sees the Hitachi ignition system consists basically
of two Kettering systems each producing an ignition pulse 180 degrees of engine
rotation apart. The system can be considered as two independent systems con-
nected together, for ignition timing, by a common breaker cam. Therefore,
two independent, identical, CDI systems could be designed to replace the
battery-coil system.

To produce a more compact, efficient, and lower cost system, it was
decided to design a system having one converter and energy storage capacitor.
Here, the circuit branched into two parts, each having its own ignition coil,
gate, and trigger circuits. The system block diagram is shown in Fig. 19.

A. DC-to-DC CONVERTER

The function of the dc-to-dc converter is to raise the low battery voltage
to an intermediate value to charge the energy storage capacitor.

Basically, the converter circuit will require more space and determine the
over-all efficiency of the circuit. To reduce its size and increase its efficiency,
the non- saturating circuit was selected over the saturating type. Generally,
however, the non- saturating circuit is more complex since it requires an ac
power source to excite the transformer. Fig. 20.

0)
O)

-3 0>

E =

-3 CP

^ z

._ o

3 Z

1- CO

II 1

SCR
Gate

> *

r >

O w

Q <»

, -r

o _

O 07

o _

.t: o

T ^

.t: o

c o

o o

c o

Q O

^
r

® 2

O) —

w O

o a.

« 5

E
o

O
O

To further reduce the size, a relatively high frequency of 10 kHz. was
selected, which permitted the use of a small powdered iron core transformer.

Originally, an intermediate voltage of 600 volts was selected. However,
an off-the-shelf inverter transformer was available that transformed the 12 V.
battery voltage to 560 V. Referring back to the computer solution, the output
voltage exceeded the 30 kV. limit, thus, the lower intermediate voltage will
lower the output voltage.

Total time between firings is 3 msec, at maximum engine speed. To
assure that arc ionization has ceased and transformer ringing decreased to
where the gate can revert to an off state, half of the 3 msec, was allotted for
the above. This leaves 1.5 msec, charging time. Since maximum bhp. is
developed at 8500 rpm. , the storage capacitor voltage was allowed to degrade
to 450 V. at 10,000 rpm.

Referring to Fig. 20, transistors Ql and Q2 act as constant current
sources until C3 has charged to about the design value. This is due to the
capacitive load on the converter transformer reflecting into the primary as a
very low impedance. Assuming a constant current charge, C3, will charge to
500 V. in 1.5 msec, if the secondary of Tl supplies 66.7 mA; this requires
a primary current of 3 A.

The astable multivibrator drives Ql and Q2 at 10 kHz. Ql and Q2 drives
Tl with 3 A. during the charging of C3. When C3 has charged to design value,
the output transistors saturate and primary current drops to the minimum
value sufficient to keep C3 charged.

>
P
o
o

o
o

I
o

d»
iZ

As mentioned earlier, the dc-to-dc converter is gated on and off. This is
accomplished by diodes D3 and D4, When terminal G is high, approximately
Vb, the diodes are in the off state, and the MV functions normally. If terminal
G is low, approximately ground potential, D3 and D4 conduct and clamp Q3 and
Q4 bases at ground potential biasing Ql and Q2 off.

Due to the low resistances in the MV circuit it has a tendency to block.
The addition of diodes Dl and D2 prevent blocking.

B. DISCHARGE CIRCUIT AND MV GATE

The schematic diagram of the discharge circuit and MV gate is shown in
Fig. 21. When a positive pulse is applied to the gate of the SCR it conducts
causing C3 to discharge through the primary of the ignition coil. Before the
SCR will revert to the off state, current must be less than the holding current.
Q5 is used to apply a turn off pulse to the converter during the time the SCR
is conducting. If the converter is not shut down during the discharge portion
of the cycle, it v^ll supply sufficient current to hold the SCR on continuously
until complete system shut-down.

The MV gate is able to completely turn off the converter over the range
of minimum holding current, 0. 5 mA. , to peak primary current of 3. 52
amperes. The gate is basically an RTL circuit with a NOR positive logic
function. Either input, when either SCR conducts, clamps the gate output to
ground potential, thus, biasing the MV off.

Since the SCR current varies over a large range, the current sensing
device should be nonlinear producing a sufficient signal at low current to

50— vV\A

o
o

c
o

— o

o
o

CVJ

ro
o

keep Q5 saturated, yet at high current the sensor output will not damage the
base-emitter junction of Q5. This requires a device having high resistance at
the holding current but low resistance at the peak primary current.

To accomplish this, diodes D9 through D12 were connected as shown. At
low currents, the diodes will not conduct unless cut-in voltage is exceeded.
The cut-in voltage for the two diodes in series is sufficient to hold Q5 at satu-
ration. The voltage drop across the diodes when peak current is applied remains
sufficiently close to the cut-in voltage to prevent excessive base current in Q5.

C. SCR TRIGGER CIRCUIT

The original breaker-plate for the Hitachi system remains as the ignition
timing source. This amplifies mechanical alterations and permits the ignition
to be readily switched back to the original battery-coil system if the CD I system
fails.

The unijunction transistor, UJT, trigger circuit is shown in Fig. 22. The
circuit produces a 5 V. peak pulse of 50 usee, duration. Referring to Fig. 23,
it is seen that the output of the trigger circuit is capable of triggering the SCR

over a wide temperature range. The SCR operates within its dissipation limits

19
and, Fig. 24, has a turn on time of 1 to 3 usee.

UJT Q6 will not conduct as long as the ignition contacts are closed. When

the contacts open, Q6 conducts and charges C4 through R12, Q6, D14, and the

low resistance of the SCR gate. After approximately 5 time constants, the pulse

applied to the SCR has returned to zero. Therefore, the values of the above

components determine the RC time constant, thus, the pulse duration. When

the points close, C4 and R14 are chosen so that the UJT can not be retriggered
r until 1 msec, has elapsed, thus preventing false triggering during the time the

points have a tendency to bounce.

Resistor R12 is selected to provide sufficient current flow to keep the
points clean, but, not to cause excessive erosion.

A complete schematic diagram of the GDI system is shown in Fig. 25.

RIO

RIA

DI4

^Tl

RI2

RI6

RI8

Rll

RI3

RI5

^^v^l ' D15

^ Q7 ,

RI7

■^12

RI9

C4,C5 = 0.1 pf
DI4,DI5=1N270
Q6,07- 2NI67I
RIO, Rll= IIOXL , 2W

RI2, RI3= II K
RI4,RI5= 500iL
RI6, RI7= 220iL
RI8, RI9= I.IK

Fig. 22. U JT Trigger.

FORWARD GATE CHARACTERISTICS

!00

6
4

SHflOEP JA INDICATES LOCUS

OF PC, ;.6LE TRIGGER'NG POINTS
_FCR V--..,;uS rCWPEHAru-»ES

>
I 6

o 2
O

-H-PrH

I ; I i

-: I ' .

irrr

—I — :-

^'21.''"^' * 'PERMITTED PULSE WIOTHS
, , y----T— FOR INDICATED "^EAK

-^ — '-t-f — - <■■ FORWARD GATE POWER

-MAXMUV GATE TRIGGER
- VOLTAGE f^C If.DICATED
^JUNCTION TE\'PERArur,E (
T,=-40*C -^ r

MAXIMUM VULPAGE AT WHICH

MAXIMUM VULFAGE AT WHICH. ^' i ■ • , , ,

■NO UNIT W.ui. 7f.:;GGch FOR I- ,- ,1 i 4 | -Ul-L.-I
Tj-^-HOO»C ' : 'I ll i I

0.001

GATE -TO -CATHODE CURRENT — AMPERES

Fie. 23.

TURN-ON TIME CHARACTERISTICS

^[|| ! i | !!l!|! i : 4 ::ritn -rH^t^H-^ - l ' n : : n i ' !

o
* 2

z I
o

^ rfPt ' i ! I ; i I

03 0.5 0.7

GATE CURRENT (Igt)-AMPERES

0.9

92CS-I3S59«1

Fig. 2k

a
u

VII. GDI SYSTEM TEST AND EVALUATION

The output of the breadboarded GDI system was measured and its charac-
teristics compared to the desired specifications listed earlier and the Hitachi
output.

The bench set-up consisted of a small battery identical to the one used for
the Hitachi ignition and a breaker plate assembly mounted on a universal motor
whose speed was variable from 1,000 to 10,000 rpm. The GDI system was
connected to spark plugs mounted on a manifold that was pressurized with nitro-
gen. Fouling was simulated by placing 500 kxi non-inductive resistors across
each plug.

Nitrogen was chosen as the pressurizing agent since the earth' s atmosphere
Is about four-fifths nitrogen and a source of nitrogen was readily available. The
use of nitrogen was not selected to simulate the conditions that exist in the com-
bustion chamber during engine operation but to give a controlled condition on
which to compare the Hitachi ignition with the GDI system.

By pressurizing the spark plug manifold with nitrogen worst case conditions
were simulated, therefore, cold plugs in a "very lean" mixture.

The following tests were performed on both ignition systems at 1, 4, and
10 thousand rpm:

1. Open circuit

a. Amplitude

b. Rise time

2. 500 ki^ load

a. Amplitude

b. Rise time

3. Spark plug impressurized

a. Amplitude of spike and table

b. Rise time

c. Table duration

4. Spark plug pressurized to 15 psi.

a. Amplitude of spke and table

b. Rise time

c. Table duration

5. Spark plug pressurized and 500 ka parallel load
. a. Amplitude of ^ite and table

b. Rise time

c. Table duration.

Figure 26 is a record of the Hitachi ignition output measurements as related
to the outlined tests. Figures 27 A through E are photographs of the output at
an engine speed of 4, 000 rpm. On examining the figures it is seen that the
ignition system failed to produce an arc for a simulated fouled plug at a high
engine rpm. If the manifold pressure was increased to 30 to 45 psi. it pro-
duced intermittent firing for test 4 and would not for any test produce an arc
at high rpm. These bench tests indicated poor ignition performance under
fouled plug conditions.

The same tests were performed with the GDI system. Results are recorded
in Fig. 28 and output photographs shown in Fig. 29 A through E.

Key system waveforms are shown in Fig. 30 A through G and depict circuit
operation. Figure 31 is a curve of power consumption as related to engine
speed. One sees that power input increases with rpm, as desired and is within
the value specified in section IV. C. 6.

The GDI system was capable of producing an arc when the plugs were
pressurized to 125 psi, where the Hitachi system failed at pressures over
40 psi. Thus the CD! system displayed superior performance over the
Hitachi system when bench tested under identical conditions.

TEST

RPM

CONDITION

1,000

4,000

10,000

Input Current (Amps)

2.67

2.3

1.7

Stored Energy (mj)

TEST I

Voltage Amplitude (kV)

Rise Time (usee)

TEST II

Voltage Amplitude

Rise Time

TEST III

Spike Amplitude (kV)

4.4

Table Amplitude (V)

500

Table Duration (usee)

1300

800

Rise Time (usee)

TEST IV

Spike Amplitude

7.2

Table Amplitude

1000

Table Duration

1000

Rise Time

TEST V

^ike Amplitude

7.2

Table Amplitude

500

Inter-

None

Table Duration

800

mittent

Rise Time

Fig. 26. Hitachi Output.

VERT = lOkV/div

ov-

HOR = 0. 5msec/div

Fig. 27A. Test I. Open circuit voltage, Hitachi System

VERT = 5kV/div

OV-

HOR = 0. 2msec/div

Fig. 27B. Test 11. 500 k load, Hitachi System.

VERT = IkV/div

OV-

HOR = 0« 5msec/ div

Fig. 27 C. Test III. Unpressurlzed Spark Plug Load.

VERT = 2kV/div

OV-

HOR = 0. Smsec/div

Fig. 27D. Test IV, Pressurized Spark Plug Load,

Hitachi System,

VERT = 2kV/div

OV-

HOR = 0. Imsec/div

Fig. 27E. Test IV. Simulated Fouled Plug,

Hitachi System.

TEST

RPM

CONDITION

1,000

4,000

10,000

Input Voltage (V)

560

425

Stored Energy (mJ)

31.4

TEST I

Voltage Amplitude (kV)

Rise Time (usee)

TEST II

Voltage Amplitude

Rise Time

TEST III

•

Spike Amplitude (kV)

5.4

Table Amplitude (V)

350

Table Duration (usee)

250

Rise Time (usee)

TEST IV

Spike Amplitude

Table Amplitude

400

Table Duration

220

Rise Time

TEST V

Spike Amplitude

Table Amplitude

500

Table Duration

220

Rise Time

Fig. 28. GDI Output.

VERT = 20kV/div

HOR = 0. 2msec/div

Fig. 29A. Test I. Open Circuit Voltage, GDI System.

VERT = 5kV/div

OV-

HOR = 0. Imsec/div

Fig. 29B. Test II. 500 k Load, GDI System.

VERT = IkV/div

OV-

HOR ^ 0. 2msec/div

Fig. 29C. Test III. Spark Plug Load, GDI System,

VERT = IkV/div

OV-

HOR = Oo 5msec/div

Fig. 29D. Test IV. Pressurized Spark Plug, GDI System.

VERT: 20V/div
Collector: Ql

VERT: IV/div
Base: Ql

HOR: 20usec/div

OV-

Fig. 30A. Converter Output Transistor Wave Form.

VERT: lOV/div

Collector: Q3

Base: Q3
VERT: 5V/div

HOR: 20usec/div

OV-

Fig. SOB. Converter Driver Transistor Wave Form.

VERT: 5V/div
SCR Gate

Points

VERT: lOV/div
HOR: 2msec/div

OV-

Fig. 30C. Trigger Circuit Wave Form.

VERT: 200V/div

C3 Voltage

Voltage Spark Plug

HOR: 2msec/div

OV-

ov-

Fig. SOD. Storage Capacitor Chaigicg Wave Form,

VERT: 20V/div
Collector Ql

CDI Output

VERT: IkV/dlv
HOR: 2msec/div

OV-

Fig. 30E. Gated Converter Drive to Inverter Transformer Wave Form.

VERT: 2V/div
SCR Cathode

CDI Output

VERT: IkV/div

HOR: 2msec/div

OV-

Fig. 30 F, MV Gate Input Wave Form.

VERT: 5V/div

Cathode D3 and D4

GDI Output

VERT: IkV/div

HOR: 2msec/div

OV-

Fig. 30G. MV Gate Output Wave Form.

«>

*-«.

— '

•o

(/)

>»

o «

Q. ^
o

; E

Vm. CONCLUSIONS AND RECOMMENDATIONS

An ignition system is required to replace the aging battery-coil ignition
now used on practically every automobile engine in use today. The Kettering
system is no longer able to meet modern automotive engine demands, and
capacitor discharge ignition systems exhibit superior performance in producing
ignition in today's high output engines.

Energy and ionization duration requirements associated with the battery
coil system cannot be applied to the CD ignition. New standards need to be
derived to best utilize the advantages of the CDI system. Brute force engineering
must be eliminated. A CDI system design must meet the conditions imposed by
the engine on which it is to be installed in order to prevent an over or under
design. Many CDI system design specifications far exceed the requirements
necessary to produce ignition.

The initial cost of a CDI system is considerably higher than that of the
Kettering. However, maintenance expenses and inconveniences imposed upon
the customer must also be considered.

The trend in industry is to cut the production cost of ignition systems,
which too often leads to systems having low reliability. To improve reliability,
initial system cost should be allowed to rise; reduced maintenance will offset
higher prices. The repayment is customer convenience, since CDI systems
can be produced having one fourth the maintenance requirements of the old
system.

Automotive history clearly refutes the belief that convenience features .
cannot compete with cost. Certainly, no one can claim that the electric starter
replaced the hand crank on a cost basis. The automatic transmission outsells
the standard transmission because it is more convenient, not less expensive.
High output eight cylinder engines are sold at a price premium in todays market
on the basis of performance, obviously not on a lower cost basis. When the
added performance and convenience can justify the added cost, the CDI system
will be accepted.

LIST OF REFERENCES

1. AC Spark Plug Service Manual , AC Spark Plug Division, General Motors
Corporation, p. 14-18, Oct. 1968.

2. Spaulding, G. E. Jr., Transistor Switched Igniton Systems , SAE Annual
Meeting, p. 11, Jan. 11-15, 1960.

3. Letter from R. E. Massoll, Spark Plug Engineering, AC Spark Plug
Division, General Motors, Subject: Ignition System Requirements,
22 Apr. 1971.

4. Champion Ignition and Engine Performance Conference , Detroit, Michigan,
1970, Automotive Technical Services Department, Champion Spark Plug
Company.

5. Champion Ignition and Engine Performance Conference , Sections I throu^
V, Toronto, Ontario, Canada, 1968, Automotive Technical Services
Department, Champion Spark Plug Company.

6. Stevens, Carlile R. , "Energy Storage and the Criteria for Proper Ignition
in the Internal Combustion Engine," IEEE Transactions on Industrial
Electronics and Control Instriunentation , vol. lECI no. 1, pg. 8-13,
Mar. 1965.

7. Engineering Manual, Champion Spark Plug Company, p. 22-23, 1966.

8. Larew, Walter B. , Igniton Systems , p. 118-201, Chilton Book Company,

9. Champion Ignition and Engine Performance Conference, Tokyo, Japan,
1969, Automotive Technical Services Department, Champion Spark Plug
Company.

10. Eason, William R. , Voltage Risetime — A New Ignition Criterion , Society
of Automotive Engineers, Automotive Engineering Congress, Detroit,
Mich. , Jan. 14-18, 1963.

11. Lovrenich, R. T. , "Automotive Ignition Problems Which Semiconductors
Can Solve," IEEE Transactions On Industrial Electronics and Control
Instrumentation, vol. lECI no. 2, p. 6-13, Sep. 1964.

12. Alexander, W. , and Sample, P. , "An Experimental Electronic Ignition
System," Electronic Engineering , vol. 36 no. 442, p. 813-816, Dec. 1964.

13. Shop Manual Honda CB 750 , Honda Motor Co. , Ltd. , p. 76-84, 1970.

14. Carroll, Ronald L. , "Electronic Ignition Systems," Electronics World,
vol. 77, p. 47-49, Feb. 1967."

15. Warner, R. S. , Welch, H. C. Pelligrino, P., Brown, R.J. , and O'Neill,
J. A. , "Solid State Ignition," Mechanical Engineering, vol. 88 na 10, p.
30-37, Oct. 1966,

16. Champion Ignition and Engine Performance Conference, Brussels,
Belgium, 1967, Automotive Technical Services Department, Champion
Spark Plug Company,

17. Van Houten, R. , and Schweitzer, J. C. , "A New Ignition System for Cars,"
Electronics , vol. 37 no. 26, p. 68-72, Dec. 5, 1964.

18. Ward, B. , "SCR Automotive Ignition System," Electronics World , vol. 72,
p. 44-45, Nov. 1964.

19. Specification Sheet on RCA 2N4103 SCR, file no. 116, Radio Corporation
of America, May 1966.

INITIAL DISTRIBUTION LIST

No, copies

1. Defense Documentation Center 2
Cameron Station

Alexandria, Virginia 22314

2. Library, Code 0212 2
Naval Postgraduate School

Monterey, California 93940

3. Assist. Professor Richard W. Adler, Code 52Ab 2
Department of Electrical Engineering

Naval Postgraduate School
Monterey, California 93940

4. LT Terrence L. Williamson, USNR 2
513 Coates Avenue

Layton, Utah 84041

5. Mr. Arthur Hoffman 1
Magnetic Circuit Elements

Henderson Way

Monterey, California 93940

6. Mr. Co Robert Smith 1
Champion Spark Plug Company

P. O. Box 910
Toledo, Ohio 43601

7. Mr. Richard E. Massoll 1
AC Spark Plug Division

GM Corp. Dept. 32-01
1300N. Dort Highway
Flint, Michigan 48556

8. American Honda Motor Co. , Inc. 1
100 W. Alondra

Gardena, California 90247

9. International Four Owners Association
1380 Garnet Avenue

San Diego, California 92109

10. Assoc. Professor Milton L. Wilcox
Code 52Wx

Department of Electrical Engineering
Naval Postgraduate School
Monterey, California 93940

11. Prof. P. F. Pucci
Code 59Pc

Department of Mechanical Engineering
Naval Postgraduate School
Monterey, California 93940

Unclassified

S<>L(iritv Cld'isifiration

DOCUMENT CONTROL DATA -R&D

\Securitv classificalion ol lille, hodv ol ab^tracl and indexing ennolfilion niu.sf be entered when the overall report /s classified)

1 ORIGINATING ACTIVITY ( Corpora le author )

Naval Postgraduate School
Monterey, California 93940

2«. REPORT SECURITY CLASSIFICATION

Unclassified

2b. GROUP

3 REPOR T TITLE

IGNITION SYSTEM REQUIREMENTS AND THEIR APPLICATION

TO THE
DESIGN OF CAPACITOR DISCHARGE IGNITION SYSTEMS

4 DESCRIPTIVE NOTES (Type o/ report and, inclusive dates)

Master's Thesis; December 1971

5 auThOR(S) (First name, middle initial, Imal rtame)

Terrence Lyle Williamson
Lieutenant, United States Naval Reserve

« REPOR T D A TE

December 1971

7«. TOTAL NO. OF PAGES

76. NO. OF RE FS

•a. CONTRACT OR GRANT NO.

6. PROJEC T NO.

9a. ORIGINATOR'S REPORT NUMBER(S)

96. OTHER REPORT NO(S) (Any other nuatbera that may be aaalgnad
this report)

10 DISTRIBUTION STATEMENT

Approved for public release; distribution unlimited.

II. SUPPLEMENTARY NOTES

12. SPONSORING MILITARY ACTIVITY

Naval Postgraduate School
Monterey, California 93940

13. ABSTR AC T

Kettering ignition systems used on the majority of automotive engines can no longer
assure reliable ignition for high-output engines. The capacitor discharge ignition, CDI,
is a promising system to supersede the obsolete battery-coil.

This study examines wave-front requirements at the spark plug for producing
ignition in the internal combustion engine and system characteristics necessary for
producing the wave-front. Arc requirements are described and used to define CDI
parameters.

A modified Kettering system which violates some basic ignition concepts was
replaced by a CDI system designed in this study. Its characteristics were derived
from arc requirements, not by aggrandizing the replaced battery-coil parameters.

During performance tests, the CDI system exhibited superior performance. It
fired simulated fouled plugs and continued to produce an arc when pressurized to 3
times the value at which the Kettering ceased to function. This improved performance
was accomplished with approximately the same stored energ}' and less input power.

DD,r:..1473

S/N 0101 -807-681 1

(PAGE 1 )

Unclassified

Security Classification

A-ai40S

Jnclassified

Security Classification

KEY wo ROS

Ignition Systems
Capacitive Discharge

BOLE WT

l.rv".,1473 'back:

FORM

I NO V 65

101 -807-682 1

Unclassified

Security Classification

A - 3 I 409

i'^2,
-^^i/^

? ?

? -r <5 5

^339

^Ppitiyt, jyst^n,

Thesis

W6273

c.l

133913

Wi 1 1 iamson

Ignition system re-
quirements and their
application to the de-
sign of capacitor dis-
charge ignition systems.

thesW6273

'SS.V^'""' requirements and their a

3 2768 000 98723 4

n DUDLEY KNOX LIBRARY